On 7 May 2026, the European Parliament and the Council of the European Union reached a provisional political agreement under the European Commission's Digital Omnibus package, amending Regulation (EU) 2024/1689 (the EU AI Act). The agreement is at the provisional stage and requires formal adoption by both institutions, expected before 2 August 2026. The urgency arose because the existing text of Article 6(2) would have required providers of stand-alone high-risk AI systems listed in Annex III to comply from 2 August 2026.
The agreement amends Articles 6(1) and 6(2) of Regulation (EU) 2024/1689. Under the revised schedule, Article 6(2) obligations, which cover Annex III high-risk AI systems, apply from 2 December 2027, extended from the original date of 2 August 2026. Article 6(1) obligations, which cover AI systems used as safety components in Annex I products subject to third-party conformity assessment, apply from 2 August 2028, extended from 2 August 2027. The agreement also introduces a new prohibited practice under Article 5 covering AI systems that generate non-consensual sexual or intimate content, effective from 2 December 2026.
Providers developing Annex III AI systems, which include tools used in employment decisions, credit scoring, biometric categorisation, and access to education or essential services, now have until 2 December 2027 to achieve conformity. Manufacturers embedding AI in safety-critical Annex I products, including medical devices, machinery, and aviation components, have until 2 August 2028. AI platforms and content generation services must remove or disable non-consensual intimate imagery generation capabilities by 2 December 2026.
The agreement also reduces from six months to three months the grace period providers of general-purpose AI models had to implement transparency markings for AI-generated content, setting the effective date at 2 December 2026. National competent authorities must establish AI regulatory sandboxes by 2 August 2027. The precise text of the amending regulation will be published after legal-linguistic review by both institutions; its exact wording may affect how providers interpret their revised compliance timelines.
Licentium advises clients active in the EU AI Act compliance cycle and can assist with high-risk AI classification, conformity assessment structuring, and preparing documentation aligned with the revised Annex III and Annex I deadlines. Our partner network includes counsel in Germany, France, and Malta. Work we undertake includes AI Act compliance scoping, provider and deployer obligation mapping, general-purpose AI model transparency obligations, and prohibited practices gap analysis.